Modeling the temporal and spatial land use change using a cellular automata-markov model

Changes in land use significantly affect the forest ecosystem, especially in terms of carbon sequestration, water quality, global climate change, and biodiversity. Therefore, determining the land use change that occurs in a certain period and revealing the land use in the coming years is essential for decision-makers and the sustainability of the ecosystem. This study was conducted in Kastamonu Regional Directorate of Forestry, Taşköprü Forest Enterprise. The primary purpose of the study is to reveal land use changes with Landsat satellite images and Geographic Information System of 1999, 2009, and 2019 and to predict the land use for 2029 using the Cellular Automato (CA-Markov) model. Land use maps for 1999, 2009, and 2019 were obtained according to the supervised classification method using Landsat 5 TM, Landsat 7 ETM+, and Landsat 8 OLI satellite images. Considering the land use maps obtained from Landsat satellite images of 1999 and 2009, the land use map 2019 was estimated with a 10-year simulation performed with the CA-Markov module. The validation and accuracy of the method were tested by comparing the simulated 2019 land use map with the 2019 land use map obtained from the Landsat satellite image. Land use map prediction for 2029 was made with the CA-Markov model. Kappa values of 0.81-1.0 of three different Landsat satellite images for which supervised classification was made in the study show that the classification was made in perfect harmony. When the 20-year land use change between 1999 and 2019 was investigated, it was revealed that the forest area decreased from 131845.4 ha to 116020.8 ha, the productive forest area decreased from 109495.2 ha to 87752.8 ha, and the broadleaf forest area decreased from 32942.3 ha to 2010.1 ha. When the estimated land use map for 2029 was examined, it was determined that the productive forest area would reduce by 0.2% (191.3 ha), the degraded forest area would decrease by 19.5% (5511.8 ha), and the total forest area would decrease by 4.9% (5703.1 ha) in the ten years. This study revealed that the CA-Markov model can effectively predict land use for future years by integrating Geographical Information System and Remote Sensing.